Re: [media-types] WG: Still need an editor....

["Murray S. Kucherawy" <[e-mail]>]

On Wed, Feb 23, 2022 at 6:20 PM Martin J. Dürst <[e-mail]>
wrote:

> I have started working on such a draft, mostly on getting up to speed
> with the newest version of xml2rfc and related tooling.
>
> When I started looking for status, update relationships, and references
> that I might have to include, I found
> https://datatracker.ietf.org/doc/html/rfc6838#section-4.2.7, which says:
>
> 4.2.7.  Additional Top-Level Types
>
>     In some cases, a new media type may not "fit" under any currently
>     defined top-level type names.  Such cases are expected to be quite
>     rare.  However, if such a case does arise, a new type name can be
>     defined to accommodate it.  Definition of a new top-level type name
>     MUST be done via a Standards Track RFC; no other mechanism can be
>     used to define additional type names.
>
> If you compare this text with what Harald and I wrote above and below,
> there's really not much (if indeed anything) that isn't already said.
>
> So before I continue working on this draft (which I can do this
> weekend), I really want to make sure that we need such a draft.
>

Hi Martin, thanks for taking this up.

RFC 6838 Sections 4.2.5 and 4.2.7 are mushy in that the former says we
should use "application" as the top-level type for anything that doesn't
clearly fit into some other top-level type, while the latter says you can
also make new top-level types in the rare case that it's necessary.  When
the "haptics" application appeared, the community had that discussion, and
it was noted that there's little guidance (only what you cited) about when
one should do one or the other.  The only other example is RFC 8081, where
a top-level type was indeed registered.

What I think we were hoping for is an update to RFC 6838 that provides some
consensus guidance about (a) when it's appropriate to make a new top-level
type versus using "application" (and/or maybe some examples of when doing
so is not appropriate); and (b) what a document declaring a new top-level
type should be required to contain.  It's fine to use RFC 8081 as a model
if people think that worked out well, or well enough to serve as a basis
for something going forward.

-MSK